如何在 Avfoundation 中正确更改采样率

Question

I have done this simple program. what it does is it just record and play back the buffers simultaneously. Everything works fine if the sample rate is 44100 hz but if I change the sample rate to 16000 or 8000, it doesn't producing any sound at all or may be some white noise which is not audiable.Why is this happening?

How can I record with different sample rate?

Following code I have tried:

import UIKit
import AVFoundation

class ViewController: UIViewController  {

var engine = AVAudioEngine()
let player = AVAudioPlayerNode()
let audioSession = AVAudioSession.sharedInstance()
let newSrc:UnsafeMutablePointer<Float>! = nil
override func viewDidLoad() {
super.viewDidLoad()



let audioSession = AVAudioSession.sharedInstance()
print(audioSession.sampleRate) // here it prints 44100 hz. because it still using the internal mic.
do {

    try audioSession.setCategory(AVAudioSessionCategoryPlayAndRecord, with: .allowBluetooth)
    try audioSession.setMode(AVAudioSessionModeDefault)
    try audioSession.setActive(true)

} catch {
}
print(audioSession.sampleRate) // here it will print 16000 hz if my bluetooth earbuds is connected, if not it will be 44100 hz.

let input = engine.inputNode
let bus = 0
let mixer = AVAudioMixerNode() // creating mixer as it is needed to set audio format

engine.attach(mixer)
engine.attach(player)
engine.connect(input, to: mixer, format: input.outputFormat(forBus: 0))

let inputFormat = input.inputFormat(forBus: bus)

engine.connect(player, to: engine.mainMixerNode, format: input.inputFormat(forBus: 0))

let fmt = AVAudioFormat(commonFormat: .pcmFormatFloat32, sampleRate: 44100.0, channels: 1, interleaved: false)

mixer.installTap(onBus: bus, bufferSize: 1024, format: fmt) { (buffer, time) -> Void in

    print(buffer.format)
    print(buffer.floatChannelData)
    print(buffer.format.streamDescription.pointee.mBytesPerFrame)
    self.player.scheduleBuffer(buffer)
    if self.player.isPlaying {
        print("true")
    }
}


engine.prepare()
do{
    try! engine.start()
    player.play()
} catch {
    print(error)
}
}

}

Answer 1

As discussed here , neither AVAudioEngine mixer nodes nor taps will do rate conversion for you. In fact in your case instead of throwing or logging an error, the mixer tap silently (get it?) gives you silence.

Since you can't use an AVAudioMixerNode for rate conversion, you can replace it with the convenient AVAudioConverter , making sure to set the correct output format of the AVAudioPlayerNode because

When playing buffers, there is an implicit assumption that the buffers are at the same sample rate as the node's output format.

If you don't do this you may hear gaps and/or pitch shifted audio.

Like so

let input = engine.inputNode
let bus = 0
let inputFormat = input.inputFormat(forBus: bus)

engine.attach(player)

let fmt = AVAudioFormat(commonFormat: .pcmFormatFloat32, sampleRate: 8000, channels: 1, interleaved: false)!
engine.connect(player, to: engine.mainMixerNode, format: fmt)

let converter = AVAudioConverter(from: inputFormat, to: fmt)!

input.installTap(onBus: bus, bufferSize: 1024, format: inputFormat) { (buffer, time) -> Void in
    let inputCallback: AVAudioConverterInputBlock = { inNumPackets, outStatus in
        outStatus.pointee = AVAudioConverterInputStatus.haveData
        return buffer
    }

    let convertedBuffer = AVAudioPCMBuffer(pcmFormat: fmt, frameCapacity: AVAudioFrameCount(fmt.sampleRate) * buffer.frameLength / AVAudioFrameCount(buffer.format.sampleRate))!

    var error: NSError? = nil
    let status = converter.convert(to: convertedBuffer, error: &error, withInputFrom: inputCallback)
    assert(status != .error)

    print(convertedBuffer.format)
    print(convertedBuffer.floatChannelData)
    print(convertedBuffer.format.streamDescription.pointee.mBytesPerFrame)
    self.player.scheduleBuffer(convertedBuffer)
}

Answer 2

This solution worked for me

let fmt = AVAudioFormat(commonFormat: .pcmFormatFloat32, sampleRate: 44100, channels: 1, interleaved: false)!
    inputNode.installTap(onBus: 0, bufferSize: 1024, format: fmt) { (buffer: AVAudioPCMBuffer, when: AVAudioTime) in
        self.recognitionRequest?.append(buffer)
    }